And slow down that loop with an if stmt to quit early if atoms overlap.
Bill
On 7/8/17 9:01 PM, Bill Ross wrote:
> Instead of an energy function, you could add up all the atom pair distances.
>
> Bill
>
>
> On 7/8/17 9:00 PM, Bill Ross wrote:
>> Hi Guqin,
>>
>> Maybe it could be a python script. E.g. start with the first molecule's
>> position as the reference. Then somehow for each other molecule
>> recursively choose a translation (or none) from the neighboring images,
>> finally when all molecules have been recursively assigned, calculate the
>> energy with an available? amber call, storing the min config seen in a
>> global variable. It could be parallelizable. Ok time-wise for a few
>> molecules, but I wouldn't want to do it with the waters in a bath.
>>
>> Bill
>>
>>
>>
>> On 7/8/17 8:34 PM, Guqin Shi wrote:
>>> Hi Bill,
>>>
>>> Could you explain a little bit more about "image minimizer"...my
>>> understanding is to write a small script that try all different orders of
>>> center-image...?
>>>
>>> Best,
>>> Guqin
>>>
>>> On Sat, Jul 8, 2017 at 10:22 PM, Bill Ross <ross.cgl.ucsf.edu> wrote:
>>>
>>>> What's needed is an image minimizer that tries all neighboring box
>>>> placements for all molecules.
>>>>
>>>> Bill
>>>>
>>>>
>>>> On 7/8/17 7:06 PM, Guqin Shi wrote:
>>>>> Hi Dan,
>>>>>
>>>>> Thanks for reply.
>>>>> Yes, the first I've thought of is to use "autoimage" directly. I have to
>>>>> admit it worked unexpectedly well. But for the late ns trajectories, it
>>>>> failed, too. It has the same result as the one I got from the separate
>>>>> center-and-image...
>>>>>
>>>>> Best,
>>>>> Guqin
>>>>>
>>>>> On Sat, Jul 8, 2017 at 8:02 PM, Daniel Roe <daniel.r.roe.gmail.com>
>>>> wrote:
>>>>>> Hi,
>>>>>>
>>>>>> Try the 'autoimage' command in cpptraj and see if that works for you.
>>>>>>
>>>>>> -Dan
>>>>>>
>>>>>> On Sat, Jul 8, 2017 at 6:59 PM Guqin Shi <shi.293.osu.edu> wrote:
>>>>>>
>>>>>>> Dear all,
>>>>>>>
>>>>>>> My simulation is on a huge complex containing 6 molecules. To visualize
>>>>>> the
>>>>>>> trajectory and get "correct" coordinates, I usually center-n-image the
>>>>>>> coordinates... I used to center on 1 molecule, image, and then center
>>>> on
>>>>>> 2
>>>>>>> molecules (include the first one), and then image; and eventually
>>>> center
>>>>>> on
>>>>>>> 6 molecules and image, as suggested in some amber mailing list posts.
>>>> It
>>>>>>> worked well.
>>>>>>>
>>>>>>> ex.
>>>>>>> center :1-300 origin (molecule 1)
>>>>>>> image center origin
>>>>>>> center :1-500 origin (molecule 2)
>>>>>>> image center origin
>>>>>>> center :1-650 origin (molecule 3)
>>>>>>> image center origin
>>>>>>> center :1-950 origin (molecule 3)
>>>>>>> image center origin
>>>>>>> center :1-1150 origin (molecule 4)
>>>>>>> image center origin
>>>>>>> center :1-1300 origin (molecule 6)
>>>>>>> image center origin
>>>>>>>
>>>>>>> However, as the molecules deviates during the simulation, after
>>>> applying
>>>>>>> the above commands, there is always one molecule (it could be different
>>>>>>> molecule dependent on the order I center-image on different masks...)
>>>>>> that
>>>>>>> just couldn't be placed back... no matter how different combinations I
>>>>>> have
>>>>>>> tried...
>>>>>>>
>>>>>>> My first question is that does it make sense that the above commands
>>>> work
>>>>>>> well for first quite a few ns of trajectories but failed to wrap for
>>>> the
>>>>>>> late ns of trajectories...?
>>>>>>>
>>>>>>> The second one will be if there is any tips or suggestions or starting
>>>>>>> point that I could try out? I have tried different combinations,
>>>> starting
>>>>>>> center on different molecule(s), I just couldn't get all 6 back to the
>>>>>>> correct place...
>>>>>>>
>>>>>>> I also tried to center on reference structure...but it just put
>>>> molecule
>>>>>>> onto the geometric center, the direction is still off...I don't know
>>>> how
>>>>>> to
>>>>>>> center molecules onto a suitable place so that all 6 could be fit into
>>>>>> one
>>>>>>> box...That doesn't make sense...I mean I start with an intact
>>>> complex...
>>>>>>> Thank you for any thoughts!
>>>>>>> Best,
>>>>>>> Guqin
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Guqin SHI
>>>>>>> PhD Candidate in Medicinal Chemistry and Pharmacognosy
>>>>>>> College of Pharmacy
>>>>>>> The Ohio State University
>>>>>>> Columbus, OH, 43210
>>>>>>> _______________________________________________
>>>>>>> AMBER mailing list
>>>>>>> AMBER.ambermd.org
>>>>>>> http://lists.ambermd.org/mailman/listinfo/amber
>>>>>>>
>>>>>> --
>>>>>> -------------------------
>>>>>> Daniel R. Roe
>>>>>> Laboratory of Computational Biology
>>>>>> National Institutes of Health, NHLBI
>>>>>> 5635 Fishers Ln, Rm T900
>>>>>> Rockville MD, 20852
>>>>>> https://www.lobos.nih.gov/lcb
>>>>>> _______________________________________________
>>>>>> AMBER mailing list
>>>>>> AMBER.ambermd.org
>>>>>> http://lists.ambermd.org/mailman/listinfo/amber
>>>>>>
>>>> _______________________________________________
>>>> AMBER mailing list
>>>> AMBER.ambermd.org
>>>> http://lists.ambermd.org/mailman/listinfo/amber
>>>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Jul 09 2017 - 00:00:03 PDT
